Lecture 5

Question 1

What is ANOVA and what purpose does it serve?

Answer 1

Analysis of Variance. Allows you to compare ANY number of means.

Question 2

What rules must be met for ANOVA?

Answer 2

1. The samples must all be random or unbiased.
2. All samples must be independent of each other.
3. Observations within each sample must be independent.
4. All of the samples must be drawn from approximately normal populations. If the data in one or more samples are severely non-normal ANOVA cannot be used unless you elect to transform all the data.
5. All samples should have approximately equal variances.

Question 3

What unique circumstance exists for equal variance for ANOVA?

Answer 3

If all sample sizes are exactly the same, tests for equal variance can be ignored.

Question 4

What is a one way ANOVA?

Answer 4

Looks at group differences based one ONE variable

Question 5

What is the difference between parametric and nonparametric tests?

Answer 5

Parametric = equal variance assumed
Nonparametric = unequal variance

Question 6

In ANOVA, what does a large F indicate?

Answer 6

Small p value

Question 7

All F-distributions are skewed ____ and have _____ degrees of freedom

Answer 7

Right, two (one for the numerator and one for the denominator)

Question 8

How do you calculate df for numerator and denominator in ANOVA?

Answer 8

Numerator df = k - 1 
Denominator df = n - k 
 
k is the number of groups
n is the total number of people in the study

Question 9

What is the F table

Answer 9

Based on the NDf and DDf, will tell you the critical value needed for significance.

Question 10

What is a type I error and what is another name for it?

Is it a big deal?

Answer 10

Type I error is considering something different when it is not. aka alpha error

BIG DEAL!

Question 11

What is a type II error and what is another name for it?

Is it a big deal?

Answer 11

Type II error is considering something not different when it is. aka beta error

Meh.

Question 12

Type __ errors are considered more deleterious to society and medicine…once you open Pandoras box, its very hard to close (e.g. vaccines cause autism)

Answer 12

Type I errors are considered more deleterious to society and medicine…once you open Pandoras box, its very hard to close (e.g. vaccines cause autism)

Question 13

What is the most common multiple comparisons test? Is it conservative?

Answer 13

Bonferroni, yes

Question 14

How do you make a Bonferroni adjustment?

Answer 14

Divide your alpha value by the number of comparisons you wish to make. This is your new p value.

Question 15

What is the limit of Bonferroni multiple comparisons before it starts to lose power?

Answer 15

5

Question 16

What type of error is possible if doing too many comparisons?

Answer 16

Type II

Question 17

What test for equal variance is used with ANOVA?

Answer 17

Bartlett’s

Question 18

What are the rules for 2 way ANOVA?

Answer 18

1. Random or unbiased sample.
2. Independent observations.
3. Normality within each group level. The larger the group size, the less this assumption is required.
4. Equal variances between each group.

Question 19

What does a 2 way ANOVA test for?

Answer 19

Interactions between factors